[xondono]

One of the common complaints is the lack of accounting for the harm done by delaying a device/therapy/drug.

If $DISEASE kills 10k people/year, and $DEVICE can save 90% of them but 5% suffer $DEVICE related deaths, delaying the release 10 years trying to get to 0% related deaths doesn’t sound like the best strategy.

[bsder]

However, this isn't just the FDA.

If $DISEASE kills 10k people/year, the company manufacturing the device doesn't have to pay anything for them.

If $DEVICE kills one child, the company manufacturing the device is on the hook for several million dollars.

So your $DEVICE needs to clear (roughly) 10K*5%*$2 million per person killed or roughly $1 billion in profit to be economically feasible.

That means that those 10k people need to spend, $100K+ each for that device to make it feasible.

It's not just the FDA that makes companies want to drive that 5% down to damn near zero.